Sumitomo Life signals a shift: digital and generative AI to support independent sales agents
In Bellevue, Washington, Yukinori Takada, president of Sumitomo Life and chairperson of the Life Insurance Association of Japan, said the company will consider using digital technologies, including generative AI, to support its sales agents. The move follows a series of scandals in the industry that exposed risks around seconded employees and called for a reset in how insurers work with agents.
"We are currently in a major turning point" for building appropriate relationships with sales agents, Takada said. The focus: support agents while protecting customer data, compliance, and independence.
What changed
- For years, insurers in Japan dispatched employees to sales agents (including banks) to help sell policies.
- Multiple scandals surfaced-data leaks tied to seconded staff and excessive favors to agents-triggering scrutiny of the practice.
- In September, the industry association issued guidelines urging companies to discontinue sending employees on loan in principle, pushing a review of insurer-agent relationships.
How AI enters the picture
Takada said Sumitomo Life has already improved operational efficiency with tools like OpenAI's ChatGPT. He emphasized that the objective now is to increase the value delivered to customers through better tools, workflow upgrades, and more effective agent support.
Sumitomo Life is looking to share digital expertise across the group. Singapore Life is noted for its tech-enabled agent management, and Sumitomo Life aims to apply those practices. The latest two-day meeting with leaders from Singapore Life and Symetra Financial focused on using AI to enhance corporate value.
Practical use cases for sales agents and customer support
- On-demand sales prep: client summaries, suitability prompts, and objection handling notes drawn from policy docs and prior interactions.
- Compliant communications: pre-approved email/text templates with disclosures and audit trails.
- Quote comparisons and policy fit: structured prompts to surface relevant riders, exclusions, and affordability checks.
- Real-time knowledge: conversational access to product rules, underwriting guidelines, and process FAQs.
- Post-call workflows: automated summaries, task capture, and CRM updates with permissioned data.
- Training and onboarding: micro-lessons, scenario practice, and quick reference guides for new products.
What leaders should do next
- Define the boundary: clarify what agent support looks like without secondments-self-service tools, virtual help desks, and accountable workflows.
- Start with a narrow AI pilot: 1-2 use cases (e.g., compliant email drafting, call summaries) for one channel or region.
- Build a small governance loop: legal, compliance, IT, and field leaders reviewing prompts, outputs, and escalations weekly.
- Use private, permissioned data: restrict sensitive fields; log prompts and outputs for audits.
- Train the front line: simple playbooks, policy-linked prompts, and clear "when to escalate to a human" rules.
- Prove ROI first: track time saved per policy, error rates, NPS, and conversion lift before scaling.
KPIs to watch
- Sales: conversion rate, average time-to-bind, cross-sell/upsell per contact.
- Service: first contact resolution, average handle time, abandonment rate, CSAT/NPS.
- Quality and risk: disclosure completeness, documentation accuracy, privacy incidents, audit findings.
- Productivity: hours saved per agent per week, turnaround time for quotes and endorsements.
Risk controls that keep you compliant
- Guardrails: block sensitive PII in prompts; use allowlisted prompts and templates.
- Human review: mandatory checks for advice, suitability, and claims-related outputs.
- Source of truth: connect AI to approved product libraries and policy documentation only.
- Auditability: log all interactions; version prompts; archive outputs in the CRM.
- Training: refreshers on disclosure, consent, and record-keeping; scenario tests for edge cases.
Group synergies and the 2026 plan
Sumitomo Life plans to highlight group-wide synergies in AI and asset management in its medium-term plan from fiscal 2026. Expect shared playbooks, common tooling, and cross-company training informed by Singapore Life's digital operations and Symetra's market experience.
The goal is clear: deliver better customer outcomes, support agent independence, and remove operational friction. AI is central to this shift, but governance and practical training will make it stick.
Upskill your team
If your sales or support teams need structured upskilling on AI workflows and prompts, explore AI courses by job for actionable, role-specific learning.
Your membership also unlocks:
